Expert System for Local Area Network Troubleshooting
Chapter One
AIMS AND OBJECTIVES
The major objectives of this project are to provide for the following:
- An expert system in the form of a software package, which can run on any system in the computer network and assist in the troubleshooting and maintenance of the LAN.
- The software package to serve as a guide to its users in carrying out troubleshooting and maintenance functions through the help of its step by step instructions.
- An expert system that can function as a tutor/trainer for getting knowledge and insight into the operation of networks (especially LANs)]
CHAPTER TWO
LITERATURE REVIEW
This chapter is concerned with a review of related literature and other research work done that are related to this project work. First, it reviews other related works done on same subject, then explains the principles involved in this work and elaborates on the various technologies available and methods of application.
REVIEW OF OTHER RELATED WORKS
Some examples of Expert Systems developed and available today include:
- Expert System for Teaching Fault Analysis
Application of expert system has been used by lecturers to teach students on subject relating to fault analysis. By using this expert system, lecturers aim to make learning more productive and efficient without necessarily increasing the staff number. The Intelligent Tutoring System (ITS) was developed by using Leonardo expert system shell. The Leonardo expert system shell is an object oriented tool for built specifically for development expert system applications and is a very useful tool for teaching fault analysis in power systems. The system was installed in a computer network in October 1994, and has been tested and implemented today. It has been found that network delivery of computer-based tutorials is the most cost effective [2].
HDDE Expert System
This ES was developed as a fault diagnostic tool for Heavy Duty Diesel Engines (HDDEs), being a Masters Degree project work done on same topic by Nabende Peter [3]. It employed Bayesian Network technology to represent the domain knowledge and reasoning for different network models relating to different HDDE fault. For each model, a Bayesian Belief Network was graphically developed representing all the possible causes for such fault, remedial actions for such fault, and associated questions to aid the troubleshooting process. The ES was built from these models using an expert system shell – Dezide Advisor [4], an advanced easy to use authoring tool and troubleshooter built upon Bayesian Network technology. Bayesian networks employ probability rules in its inferencing system. A user interface for the ES was developed with the aid of C-Sharp programming language. The user then selects a particular problem associated with a HDDE through the interface which then opens up the Troubleshooter tool, from which one can then select the appropriate model to make diagnoses.
- ORBI
This expert system for Environmental Impact Assessment (EIA) was developed by the Universidade Nove de Lisboa for the Portuguese Department of Environment [5]. It is written in Prolog with a natural-language interface (Portuguese). This is one of the few expert systems for Environmental Impact Assessment which currently has such a facility. This facility makes communication with the system more straight-forward and easier than with the usual expert system shells or programming languages, as the user can program and update the system in the user’s own language rather than using computer code. The system consists of four separate programs: the first is in BASIC for digitizing a subject region (a region is broken down into a locational grid); the second is for receiving and storing the digitized information; the third is for knowledge updating; and the fourth comprises two modules in Prolog (the first is a natural language interface, and the second is an evaluation and explanation module; each of which runs independently of the other) [6].
ORBI has expertise in several disciplines, including geology, hydrology, ecology, and microclimate. The system makes judgments about the suitability of a particular subject region for uses such as industry, agriculture, and recreation. Quantified values are obtained via inferencing rules that indicate the suitability of the environmental assets available relative to development requirements (for example, intensive agriculture). ORBI can produce graphic output in the form of maps, via a plotter.
CHAPTER THREE
METHODOLOGY, SYSTEM DESIGN AND IMPLEMENTATION
RESEARCH METHODOLOGY
Before commencing the process of developing this expert system, some research had been done regarding the previous and current technology of the Expert System. The research reviewed various literatures in order to understand the concepts behind the development of the expert system including the critical elements involved in the development and the technologies needed. The phase also involved doing research on related work and identification of gaps in the techniques and tools that are currently available. This phase was important in that it involved knowing how well expert systems perform in their ability of assisting and replacing the human expert as well as the user. The result of research and reviews helped much in giving the idea and insight into the existing application of expert system to be implemented in this system.
KNOWLEDGE ACQUISITION AND ANALYSIS
Expert system development is all about applying human expertise into computer verse, which is based greatly on the integration of human knowledge in the system. Thus knowledge acquisition is the heart of the expert system. After reviewing the outcomes of the literature research, the fundamental and basic concept of expert system was identified in order to guide the determination of the problem’s domain. The appropriateness of the problem was taken into consideration as to make sure that it is suitable to be solved by an expert system. After the problem domain was determined, the knowledge and information based on the problem was then acquired. The knowledge sources were ascertained which were the experts in the domain itself. The collection of knowledge for the ‘Quickhelp Troubleshooter Expert System’ – the name given to the finished product – was done by interviewing those who are experienced in the field of computer networking, and collection of data from such related maintenance and repair manuals. Other information was obtained from the internet, and various other books relating to the field.
CHAPTER FOUR
SYSTEM INSTALLATION AND TESTING
Implementation involved developing graphical user interfaces that introduce the expert system and guides the user in selecting the particular Local Area Network malfunction to diagnose. These graphical user interfaces were developed using Visual Basic programming language.
CHAPTER FIVE
CONCLUSION AND RECOMMENDATIONS
CONCLUSION
In this project, an Expert System for diagnosing and repairs of Local Area Network malfunctions has been developed – herein called Quickhelp Troubleshooter.
At first, different possible LAN malfunctions were documented through the knowledge acquisition process. These were carefully represented and modeled so as to provide for proper development of the ES knowledge base. The developed Expert system was tested and validated against the diagnostic process of experienced technicians who are involved in the maintenance and repair of Local Area Networks and the troubleshooting procedures recommended in repair manuals. Results showed that the Expert System made diagnoses similar to those of the experienced technicians. This demonstrated successful implementation of the Expert System.
The expert system has a simple easy to use interface such that users who are not very knowledgeable with computer software will find it easy to operate. This simplicity comes in especially in the structure of the steps performed during diagnosis, since in most cases the user has to answer with a ‘Yes’ or ‘No’; and perform any necessary action recommended by the software. The expert system can also be used as a training tool for new or inexperienced network technicians.
RECOMMENDATIONS
The Expert System that has been developed can still be improved upon. The following are recommended as further workThe automation of updating the knowledge base and reasoning steps when new information is gained from experience obtained after using the Expert System. The scope for this work can also be extended such that it does not only cover Local Area Networks (LAN), but Metropolitan Area Networks (MAN) and Wide Area Networks (WAN); since most networks in reality are large.
In this particular work, computers and computerized devices were considered as the nodes or devices/equipment within the network. In real life situations, networks can be implemented using telephone networks (e.g in PABX or intercom systems), television systems (e.g in CCTV), and other devices. This work can be extended to cover such other network devices.
The Expert System can have the capability to be used in different languages. Further work is recommended so that it can have the capability of being used in local languages.
REFERENCES
- Forouzan B. A., (2007). Data Communications and Networking, McGraw Hill, Singapore, 4thedition, pp 7, 372
- Negnevitsky M., (1998). A knowledge based tutoring system forteaching fault analysis, Power Systems, IEEE Transactions on Volume 13, Issue 1, Feb. 1998 pps 40 – 45
- Nabende P., (2006). An Expert System for Diagnosing Heavy DutyDiesel Engine faults. http://rug.academia.edu/PeterNabende/Papers
- Dezide ApS. (2001 – 2005). Dezide Advisor User Guide. http://www.dezide.com
- Geraghty, P. J., (1993). Environmental assessment and the applicationof expert systems: an overview. J. Environ. Manage. 39: 37-38.
- Asian Development Bank, (1997). Environmental Impact Assessmentfor Developing Countries in Asia: Volume 1. adb.org/Documents/Books/Environment_Impact/chap8.pdf
